advertisement


Paralleling NCC200

LuckyLuke

Member
Hello

My first post on PFM so hi everybody!
Few years ago I bought 8 of NCC200 power amplifiers PC board kits and got great success with them.
By the way I'm talking about the old, single-sided ones.
Right now I have two left of these.
And I have a project for them.
I want to build myself an Audio Lab amp using two paralleled NCC200 with the following features.

1) The amp needs to approach an almost perfect voltage source so power needs to double as impedance is halved, right to 2 Ohms.
2) Output impedance of the parraleled amps should be below .01 ohms for at least half of the available output voltage, all this from 10 to 40KHz.
2) I don't ask for lot of power...only 50-60 watts total, considering with the two amps paralleled, which are powered from low DC voltage supply but lots of current...from 6.7 Amps AC from transformer.
So this means 20 Vrms output but it has to stay 20 V down to 2 Ohms.
3) I will get a 320VA toroidal Transformers with 48Vct/6.7 amps with generous capacitance.
4) The last amps I built with the others had a low offset error below 5mVdc and gain precision better than 0.1dB and I will surely get these specs for this project.
5) I understand I will need to couple the two amps together through a pair of 0.22 Ohms(???) resistors.

Am I on a good track?
Anything you could suggest to help?
Thanks!

Lucky Luke
 
You could just build one NCC board per channel and then add extra output devices in parallel with the existing ones. The A&R SA200 did this, as well as some other amps i'm sure. I guess you'll want to buy a big batch and hand-match them for best performance. I think you might want to use a bigger transformer too, especially with extra output devices... Maybe go dual-mono and have 2 transformers?
 
Hello SJ

Thanks for suggestion.
Well, I didn't think about using multiple output devices to increase output current capabilities.
It's a good idea and surely the most logic one!
I now understand why I chose to use two amplifiers and parallel them together and I realize now it's a bad choice as my coupling two 0.22 ohms will simply increase my output impedance.
I made a mistake that I recognise.
The reason why is the fact that my very first idea was to use two or three LM3866 ot TDA7293 and couple these trough resitors to increase available current, as did Linn with their Klimax amp.
As I remembered I had two NCC200 left I just figured I could do much better with excellent amp circuit as these NCC are and couple them together with resistors...it was a mistake.
Question : putting more output devices in parallel with the NCC200's.
Simply connect them in parallel?
Naturally increasing substantially my PSU's current output.

Luc
 
If you use BUV20 output transistors these will give you 10A output current on 30V DC rails and you won't need to parallel them - one output pair will suffice.
 
Extra output transistors can be added in paralel - they need independent emitter resistors. There is no free lunch - you have increased the capacitance load on the gain stage, so I would want to check that the stability is not compromised. It will certainly reduce high frequency performance slightly.

You need scope, dummy loads, and signal generator. Check that there is no ringing on square waves for resistive loads, and that any overshoot on capacitative loads is well damped. If there are stability problems, you will need expert advice to fix the loop compensation.

What load do you want to drive that needs the extra output current? Apogee speakers?
 
Extra output transistors can be added in paralel - they need independent emitter resistors. There is no free lunch - you have increased the capacitance load on the gain stage, so I would want to check that the stability is not compromised. It will certainly reduce high frequency performance slightly.

You need scope, dummy loads, and signal generator. Check that there is no ringing on square waves for resistive loads, and that any overshoot on capacitative loads is well damped. If there are stability problems, you will need expert advice to fix the loop compensation.

What load do you want to drive that needs the extra output current? Apogee speakers?

Hello PigletsDad

Thanks for reply.
It is great advice from you but I'm not at the testing point right now.
Nothing concrete on my test bench!
I'm planning for incoming future.
I always try to get things the rightest way before starting something.
Like life, designing should be smooth, silky...frustrationless!
The reason why I ask for advices.

You said :" you have increased the capacitance load on the gain stage".
Curious...What capacitance?

"What load do you want to drive that needs the extra output current?"

It's a Lab amp that has very low output impedance and this LowZ must be quite linear in respect of frequency.
It must approach a perfect voltage source down to at least 2 Ohms.
These two specs should be perfectly respected up to output voltage of 10 Volts.
So since amp is good for 20 Volts it should not be a problem if I use a"beefed-up" output section and PSU.
It should be used to test sometimes reactive parts but most of the time loudspeakers, components and systems.
And what interests me is the non-linear caractéristics of these.
So this kind of amp is quite welcome.

Oh by the way I have a lot of MJ21194 left with all HFE at 1% tol.


LuckyLuke
 
Each transistor has capacitance between its terminals. In this case the Base to collector capacitance is probably dominant, and presents the main high frequency load on the VAS gain stage. If you use two transistors, you get double the capacitance. Cbe is numerically larger, but mostly removed by unit gain operation - in effect it is bootstrapped away.

For parallel use, HFE is not the issue (the transistors are voltage controlled, not current controlled, due to the Miller feedback in the VAS stage), but VBE at the idle bias current. You want matching to better than a few mV.

"Approach a perfect voltage source" is meaningless - like marketing hype.

What output impedance do you need for your lab application?
 
Each transistor has capacitance between its terminals. In this case the Base to collector capacitance is probably dominant, and presents the main high frequency load on the VAS gain stage. If you use two transistors, you get double the capacitance. Cbe is numerically larger, but mostly removed by unit gain operation - in effect it is bootstrapped away.

For parallel use, HFE is not the issue (the transistors are voltage controlled, not current controlled, due to the Miller feedback in the VAS stage), but VBE at the idle bias current. You want matching to better than a few mV.

"Approach a perfect voltage source" is meaningless - like marketing hype.

What output impedance do you need for your lab application?

Hello PigletsDad and thanks for reply

"...meaningless - like marketing hype"

Yes I know perfectly what you mean but I insist on the terms almost perfect as this does not represent a "fidelity fudge factor" but simply a desirable Zout and Zfrequency linearity for the unique reason of precision steady state measurements.

And I'm reasonnable...0.01 ohms is good enough. Some call this a damping factor of 100 (Sorry but :D) but I repeat it's not for marketing reasons, just low Zout...period.

By the way this is only an unusual application of amplifier.

LuckyLuke
 
The stock version will do that; the open loop impedance is about 0.2 ohms. At low frequencies the feedback factor is about 100, falling at 6db/octave once the compensation cuts in. This means impedance will rise from a few milliohms. At the top of the audio band, I am pretty sure it will remain below 10milliohms.

For stability in audio applications, an output inductor is normally used, but in your application you may need to remove that.
 
The stock version will do that; the open loop impedance is about 0.2 ohms. At low frequencies the feedback factor is about 100, falling at 6db/octave once the compensation cuts in. This means impedance will rise from a few milliohms. At the top of the audio band, I am pretty sure it will remain below 10milliohms.

For stability in audio applications, an output inductor is normally used, but in your application you may need to remove that.


The stock version will do that? Half voltage? Down to 2 Ohms and give proportionnal current? At .01 ohms...from 10 to 40KHz?
With all respect to this amp that I like very much...I don't think so and this is not a lack of irrespect I have towards it.
Also all this is not necessarily an absolute required set of specs for fidelity.
But what I want is something else and I repeat in different term, it's not to compete with actual design or pretending building a sweeter "pitbull" amp than what NCC is but just a Lab amp with preceeding specs.

Lucky Luke
 
By the way...

df_genius talked precedently about a single 50 amps output device, a BUV20 which is unfortunately obsolete.
Maybe another type will do?
I like the ldea of having only a pair instead of four or more to avoid some pifalls and also increased capacitance load on the gain stage that you mentionned earlier.

LL
 
Yes, I was assuming BUV20 output devices.

The basic issue is the fall of Hfe with IC. If you want to be able to drive 20V over 2Ohms, you need an output device with high Hfe at 10A, either as one high current transistor, or multiple in parallel .

For example, look at http://www.onsemi.com/pub_link/Collateral/MJ21193-D.PDF

On page 3, you can see graphs of Hfe versus IC. For this device, you can see that current gain has dropped by over 20% at 10A. The local connection topology and the global feedback both mitigate this, but a device with a flatter gain profile will do better.

I guess another obvious question - if it is for a lab application, is output impedance the relevant factor, or nonlinear gain error. In some cases, linear errors can be calibrated away in a measurement scenario, but nonlinear effects dominate the error budget.
 
This site contains affiliate links for which pink fish media may be compensated.
This is an old thread, but which output devices are being currently fitted in NCC200 and NCC220 boards?
 


advertisement


Back
Top